Written by Marykate Krasner
Until recently, Facebook’s position at the heart of digital advertising seemed unshakeable. In addition to its reach—2.89 billion monthly active users globally—Facebook offered tools like Audience Insights, super-granular segmenting, and a tracking pixel that could follow people around the internet with a high degree of accuracy (and/or creepiness, depending on how you see it).
But now, Facebook Analytics and Audience Insights are gone, thanks to iOS 14 privacy updates. Even that massive user base could shrink: only 51% of U.S. teens say they use Facebook, down from 71% just a few years ago. In what may be the clearest sign of changing times, a 2020 study found that advertising on Amazon now delivers higher ROI than Facebook, by a significant margin.
Given all this, should you consider shifting ad dollars toward Amazon and away from Facebook? Yes—with a few caveats.
We knew Apple’s privacy rules would make Facebook ads less reliable and less useful than they once were. Several months in, we’re reckoning with just how much has changed. Recently, Facebook said that they are underreporting iOS web conversions by approximately 15%. Here at Part and Sum, we’ve even advised clients to stop using ROAS as a key metric—when it used to be the most important metric of all, and an easy way to measure performance. Now, it’s so unreliable that it’s become useless. Instead, businesses need to pick and choose from other metrics like conversion rate, lifetime value, and customer acquisition cost to get a sense of how they’re doing.
We’ve also found it difficult to test creative and value proposition messaging accurately on Facebook, with big discrepancies in creative data when comparing on-platform metrics to other reporting tools like Google Analytics. Facebook has also recommended analysis be done at the campaign level, rather than the creative level, since their model for attribution estimation has been estimating creative conversion incorrectly.
Big difference in reporting on creative: Facebook shows five purchases, Google Analytics shows 10.
With over 140 million active Prime subscribers in the U.S. alone, Amazon offers advertisers an opportunity to reach customers and gain the kind of insights they used to see on Facebook.
One of Amazon’s greatest strengths is the wide array of properties it owns, from e-comm sites like Shopbop and Zappos, to streaming platform Twitch and VOD service Amazon Prime Video, to in-home smart devices like Alexa, Echo, and FireTV. Plus, customers can use Amazon Pay to check out at their favorite online stores.
All of this gives Amazon access to the shopping, browsing, and viewing behavior of millions of users, and that data has allowed them to improve their ad targeting and algorithms. (In other words, yes, Alexa is listening to you.) Although Amazon’s reporting isn’t as robust as Facebook was at its peak, advertisers can still find helpful insights, if they know where to look.
Unlike Facebook, which only shows ad attributed metrics in-platform, the Retail Insights report lets advertisers see non-ad attributed metrics for their products, helping them understand the direct correlation between ad campaigns and earned performance on Amazon.
This report shares metrics like conversion rate, earned media impressions, orders prior to a campaign’s start date, orders during the campaign, and post-campaign orders, so advertisers can easily see the long-term lift from paid advertising. Retail Insights also provides competitive metrics, like subcategory growth, allowing advertisers to understand how often their ads appear relative to other advertisers in their subcategory.
Market Basket offers a rare glimpse inside customers’ shopping habits. This report shows advertisers the top three products most frequently purchased at the same time as their brand’s products: After all, most of us add more than one item to our Amazon cart before checking out.
Digging deeper, advertisers can see frequency (the percentage of time these top three products were purchased alongside their own products) and time (by selecting specific date ranges). These unique insights can help identify opportunities for bundling, cross-marketing, and perhaps even product line diversification.